76 Fa

An Open-Source Research Kit for the
da Vinci® Surgical System
Peter Kazanzides
Research Professor
Dept. of Computer Science
Johns Hopkins University
NSF Engineering Research Center for Computer Integrated Surgical Systems and Technology
Copyright © CISST ERC, 2015
Outline
• Overview of da Vinci Research Kit
• Control architecture
• Open source mechatronics
• Open source software
• Community
• Interoperability
• Future outlook
da Vinci Research Kit
• Mechanical components from da Vinci
JHU 2008
“classic” systems
• Donated by Intuitive Surgical to selected
university labs
• Consortium to provide “open source”
engineering and support
• Software – JHU (CISST/SAW)
• Controller electronics – JHU
• Interface electronics – ISI
• Controller power/packaging – WPI
• Controllers and software also adapted for
use with complete recycled da Vinci
“classic” systems
Worldwide 2014 • http://research.intusurg.com/dvrkwiki/
DVRK Video
Robot Control Architecture: Centralized
Computation and Distributed I/O
App Traj Servo
Computer IO IO IO
(multicore)
Amp Amp Amp
Power and
high-speed Robot
serial network
(e.g., Firewire)
• Possible with high-performance networks/computers

– All the benefits of distributed I/O (reduced cabling)
– Computation on familiar development platform (flexibility)
Serial Network Requirements
• Speed of at least 100 Mbits/sec
• Low latency (tens of µsec)
• Ability to daisy-chain
• Readily available
• Simple FPGA implementation ?
• Option for high-flex cabling
IEEE-1394a (FireWire)
EtherCAT
In 2006, FireWire appeared to be best choice
Today, other good options, such as EtherCAT
Open Source Mechatronics
Open source: http://jhu-cisst.github.io/mechatronics/

• Schematics, PCB layout, firmware (Verilog)
Hardware Block Diagram
IEEE 1394
PHY
+
DAC Power
Op Amp
write
SPI -
Motor
read
Differential Sense
Filter Amplifier
Amplifier Resistor
SPI Motor current

buffers ADC
Pot Power
Op Amp Pot
quad.
Filter
FPGA
Level
Translation Enc
Control Performance
• Latency primarily due to overhead on PC
– ~35 µs per asynchronous transaction
– Individual read/write to 8 boards: ~530 µs
– Taking advantage of broadcast and peer-to-peer
transfers reduces I/O to ~116 µs
Broadcast protocol, 8 board sets
Hardware Revisited
• In 2006, FireWire (IEEE-1394) seemed to be a good
choice, but less common today than Ethernet
– Most desktops and laptops contain Ethernet port
– Linux has good low-level support for FireWire (libraw1394);
other platforms do not
– Development environments, such as Simulink Real-Time
(formerly Matlab xPC) and Labview CompactRIO, support
Ethernet, but not FireWire
– Real-time driver for Ethernet available (RTnet); RT-Firewire
no longer supported
– EtherCAT uses standard Ethernet hardware on master, with
custom slave hardware to enable daisy-chaining
37
Ethernet Alternatives
FPGA- FPGA-
ecat ecat
EtherCAT
FPGA- FPGA-
ecat ecat
Ethernet FireWire
Bridge
Board
Ethernet-FireWire Bridge Performance
Measured performance on 8- Analytical performance comparison

node system (DVRK)
Ethernet-FireWire Design
da Vinci Teleoperation (with ROS)
Simulation/Visualization
ROS/rviz
Gazebo Model
Dynamic Simulation
Matlab Models
PSM & MTM
User Community
UBC
CSTAR SUNY Buffalo
WPI
Vanderbilt CMU JHU

Stanford
First DVRK User Group Meeting
Johns Hopkins University, March 20-21, 2014
Interoperability
da Vinci master teleoperating

Raven II via ROS
Community Research:
Learning by Observation
Summary
• Open source research platform based on
first-generation da Vinci system:
– Open source electronics
– Open source software, with ROS interfaces
• Growing user community:
– 20 sites up and running (internationally)
– several more in process
• Links:
– http://research.intusurg.com/dvrkwiki/
– http://github.com/jhu-cisst (and jhu-saw)
– http://github.com/jhu-dvrk
Acknowledgments
• Partners: Russell Taylor (JHU), Simon DiMaio (ISI), Greg
Fischer (WPI)
• Software: Anton Deguet, Zihan Chen, Simon Leonard, Balazs
Vagvolgyi, Jonathan Bohren, Kwang Young (Eddie) Lee, Ankur
Kapoor, Tian Xia
• Firmware: Zihan Chen, Paul Thienphrapa, Long Qian
• Hardware and Testing: Lawton Verner, Ravi Gaddipati, Gang Li,
Nirav Patel, Zhixian Zhang, Alex Camillo
• Intuitive Surgical: Simon DiMaio, Arpit Mittal, Kollin Tierling, Dale
Bergman
• Support: NSF EEC 9731748, EEC 0646678, MRI 0722943, NRI
1208540, NRI 1327657, NASA NNX10AD17A
Questions?
Stereo Viewer
da Vinci Master Tool

Manipulators (MTMs)
da Vinci Patient Side

Manipulators (PSMs)
IEEE-1394 controllers (4
Control PC
enclosures with 8 nodes) Footpedal Tray

76 Fa

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

76 Fa

Uploaded by

Copyright:

Available Formats

An Open-Source Research Kit for the

da Vinci® Surgical System

App Traj Servo

• Possible with high-performance networks/computers

Open source: http://jhu-cisst.github.io/mechatronics/

SPI Motor current

Measured performance on 8- Analytical performance comparison

CSTAR SUNY Buffalo

Vanderbilt CMU JHU

Johns Hopkins University, March 20-21, 2014

da Vinci master teleoperating

da Vinci Master Tool

da Vinci Patient Side

You might also like